2017년 7월 16일 일요일

No Texting While Driving


This chapter walks you through the development of No Texting While Driving, an app that autoresponds to text messages you receive while you're driving. The app, first created with App Inventor by a beginning computer science student, is similar to a now-mass-produced app developed by State Farm Insurance. It is a prime example of how App Inventor provides access to some of the great features of the Android phone, including SMS text processing, database management, text-to-speech, and the location sensor.



In January 2010, the National Safety Council (NSC) announced the results of a study that found that at least 28 percent of all traffic accidents-close to 1.6 million crashes every year-are caused by drivers using cell phones, and at least 200,000 of those accidents occurred while drivers were texting 

(http://www.nsc.org/pages/nscestimates16millioncrashescausedbydriversusingcellphonesandtexting.aspx). As a result, many states have banned drivers from using cell phones altogether.
Daniel Finnegan, a student in the Fall 2010 session of the University of San Francisco App Inventor programming class, came up with a great app idea to help with the driving and texting epidemic. The app he created, which is shown in Figure 4-1, responds automatically (and hands-free) to any text with a message such as "I'm driving right now, I'll contact you shortly."

Some in-class brainstorming led to a few additional features that were developed for a tutorial posted on the App Inventor site:
The user can change the response for different situations

  For example, if you’re going into a meeting or a movie instead of driving, the response can be modified accordingly.

The app speaks the text aloud
  Even if you know the app will autorespond, the jingle of incoming texts can kill you with curiosity.
The response message can contain your current location

Figure 4-1. The No Texting While Driving app

  If your partner is at home making dinner, he or she would probably like to know how much longer your commute will last, without endangering you by having you answer the text.
Some weeks after the app was posted on the App Inventor site, State Farm Insurance created an Android app called "On the Move," which has similar functionality to No Texting While Driving (http://www.statefarm.com/aboutus/newsroom/20100819.asp). The service is free to anyone, as part of State Farm's updated Pocket Agent for AndroidTM application, which the company announced in a YouTube video that can be found here: http://www.youtube.com/watch?v=3xtjzO0-Hfw.

We don't know if Daniel's app or the tutorial on the App Inventor site influenced "On the Move," but it's interesting to consider the possibility that an app created in a beginning programming course (by a creative writing student, no less!) might have inspired this mass-produced piece of software, or at least contributed to the ecosystem that brought it about. It certainly demonstrates how App Inventor has lowered the barrier of entry so that anyone with a good idea can quickly and inexpensively turn his idea into a tangible, interactive app.

What You'll Learn

This is a more complex app than those in the previous chapters, so you'll build it one piece of functionality at a time, starting with the autoresponse message. You'll learn about:

⦁ The Texting component for sending texts and processing received texts.
⦁ An input form for submitting the custom response message.
⦁ The TinyDB database component for saving the customized message even after the app is closed.
⦁ The Screen.Initialize event for loading the custom response when the app launches.
⦁ The Text-to-Speech component for speaking the texts aloud.
⦁ The LocationSensor component for reporting the driver’s current location.

Getting Started

For this app to work, you need a text-to-speech module, Text-To-Speech Extended, on your phone. This module is included in Android version 2 or higher, but if you are running an Android 1.x operating system, you'll need to download it from the Android Market. On your phone:

1. Open the Market app.
2. Search for TTS.
3. Select the app Text-To-Speech Extended to install.

Once the Text-To-Speech module is installed, open it to test its features. When it opens, set the default language as desired. Then select "Listen to Preview." If you don't hear anything, make sure the volume on your phone is turned up. You can also change the way the voice sounds by changing the setting for the TTS Default Engine property.

After you've set up the Text-To-Speech module to your liking, connect to the App Inventor website and start a new project. Name it "NoTextingWhileDriving" (project names can't have spaces) and set the screen's title to "No Texting While Driving". Once the screen title has been changed open the Blocks Editor and connect to the phone through WiFi.

Designing the Components

The user interface for the app is relatively simple: it has a label that displays the automated response, along with a text box and a button for submitting a change. You'll also need to drag in a Texting component, a TinyDB component, a TextToSpeech component, and a LocationSensor component, 
all of which will appear in the "Nonvisible components" area. You can see how this should look in the snapshot of the Component Designer shown in Figure 4-2.

Figure 4-2. The No Texting While Driving app in the Component Designer

You can build the user interface shown in Figure 4-2 by dragging out the components listed in Table 4-1.

Set the properties of the components in the following way:

⦁ Set the Text of PromptLabel to "The text below will be sent in response to all SMS texts received while this app is running."
⦁ Set the Text of ResponseLabel to "I'm driving right now, I'll contact you shortly." Check its boldness property.
⦁ Set the Text of NewResponseTextbox to " ". (This leaves the text box blank for the user's input.)
⦁ Set the Hint of NewResponseTextbox to "Enter new response text."
⦁ Set the Text of SubmitResponseButton to "Modify Response."

Adding Behaviors to the Components

You'll start by programming the basic text autoresponse behavior, and then successively add more functionality.

Table 4-1. All the components for the No Texting While Driving app

Programming an autoresponse

For the autoresponse behavior, you'll use App Inventor's Texting component. You can think of this component as a little person inside your phone that knows how to read and write texts. For reading texts, the component provides a Texting.MessageReceived event block. You can drag this block out and place blocks inside it to show what should happen when a text is received. In the case of this app, we want to automatically send back a prewritten response text.

To program the response text, you'll place Texting1.SendMessage block within the Texting1.MessageReceived block. Texting1.SendMessage actually sends the text-so you'll first need to tell the component what message to send, and who to send it to, before calling Texting1.SendMessage. Table 4-2 lists all the blocks you'll need for this autoresponse behavior, and Figure 4-3 shows how they should look in the Blocks Editor.

Table 4-2. The blocks for sending an autoresponse

Figure 4-3. Responding to an incoming text

How the blocks work

When the phone receives a text message, the Texting1.MessageReceived event is triggered. As shown in Figure 4-3, the phone number of the sender is in the argument number, and the message received is in the argument messageText. For the autoresponse, the app needs to send a text message to the sender. To send the text, you first need to set the two key properties of the Texting component: PhoneNumber and Message.TextingPhoneNumber is set to the number of the sender, and Texting.Message is set to the text you typed into ResponseLabel: “I’m driving right now, I’ll contact you shortly.” Once these are set, the app calls Texting.SendMessage to actually send the response.
Comments are a very important part of programming, they can be used to inform other programmers of important information about the code. You can add comments by right-clicking a block and selecting Add Comment. 

Once you add a comment, you can show or hide it by clicking the blue question mark that appears. You don’t have to add comments in your app-we've simply included them here to help describe each block and what it does.

Most people use comments to document how they are building their app; comments explain how the program works, but they won't make the app behave differently. Comments are important, both for you as you build the app and modify it later, and for other people who might customize it. The one thing everyone agrees on about software is that it changes and transforms often. For this reason, commenting code is very important in software engineering, and especially so with open source software like App Inventor.

Test your app. You'll need a second phone to test this behavior. If you don't have one, you can register with Google Voice or a similar service and send texts from that service to your phone. From the second phone, send a text to the phone running the app. Does the second phone receive the response text?

Entering a Custom Response

Next, let's add blocks so the user can enter her own custom response. In the Component Designer, you added a TextBox component named NewResponseTextbox; this is where the user will enter the custom response. When the user clicks on the SubmitResponseButton, you need to copy her entry (NewResponseTextbox) into the ResponseLabel, which is used to respond to texts. Table 4-3 lists the blocks you'll need for transferring a newly entered response into the ResponseLabel.

Table 4-3. Blocks for displaying the custom response

How the blocks work

Think of how a typical input form works: you first enter something in a text box, and then click a submit button to tell the system to process it. The input form for this app is no different. Figure 4-4 shows how the blocks are programmed so that when the user clicks the SubmitResponseButton, the SubmitResponseButton.Click event is triggered.

Figure 4-4. Setting the response to the user's entry

The event handler in this case copies (or sets, in programming terms) what the user has entered in NewResponseTextbox into the ResponseLabel. Recall that ResponseLabel holds the message that will be sent out in the autoresponse, so you want to be sure to place the newly entered custom message there.

Test your app.Enter a custom response and submit it, and then use the second phone to send another text to the phone running the app. Was the custom response sent?

Storing the Custom Response in a Database

You've built a great app already, with one catch: if the user enters a custom response, and then closes the app and relaunches it, the custom response will not appear (instead, the default one will). This behavior is not what your users will expect; they'll want to see the custom response when they restart the app. To make this happen, you need to store that custom response persistently.

You might think that placing data in the ResponseLabel.Text property is technically "storing" it, but the issue is that data stored in component properties is transient data. Transient data is like your short-term memory; the phone "forgets" it as soon as an app closes. If you want your app to remember something persistently, you have to transfer it from short-term memory (a component property or variable) to long-term memory (a database).

To store data persistently, you'll use the TinyDB component, which stores data in a database that's already on the Android device. TinyDB provides two functions: StoreValue and GetValue. The former allows the app to store information in the device's database, while the latter lets the app retrieve information that has already been stored.

For many apps, you'll use the following scheme:

1. Store data to the database each time the user submits a new value.
2. When the app launches, load the data from the database into a variable or property.

You'll start by modifying the SubmitResponseButton.Click event handler so that it stores the data persistently, using the blocks listed in Table 4-4.

Table 4-4. Blocks for storing the custom response with TinyDB 

How the blocks work

This app uses TinyDB to take the text it just put in ResponseLabel and store it in the database. As shown in Figure 4-5, when you store something in the database, you provide a tag with it; in this case, the tag is "responseMessage." Think of the tag as the name for the data's spot in the database; it uniquely identifies the data you are storing. As you'll see in the next section, you'll use the same tag ("responseMessage")when you load the data back in from the database.
Figure 4-5. Storing the custom response persistently

Retrieving the Custom Response When the App Opens

The reason for storing the custom response in the database is so it can be loaded back into the app the next time the user opens it. App Inventor provides a special event block that is triggered when the app opens: Screen1.Initialize (if you completed MoleMash in Chapter 3, you've seen this before). If you drag this event block out and place blocks in it, those blocks will be executed right when the app launches.
For this app, your Screen1.Initialize event handler should check to see if a custom response has been put in the database. If so, the custom response should be loaded in using the TinyDB.GetValue function. The blocks you'll need for this are shown in Table 4-5.

Table 4-5A. Blocks for loading the data back in when the app is opened

Figure 4-5B. Blocks for loading the data back in when the app is opened (continued)

How the blocks work

The blocks are shown in Figure 4-6. To understand them, you must envision a user opening the app for the first time, entering a custom response, and opening the app subsequent times. The first time the user opens the app, there won't be any custom response in the database to load, so you want to leave the default response in the ResponseLabel. On successive launches, you want to load the previously stored custom response from the database and place it in the ResponseLabel.
Figure 4-6. Loading the custom response from the database upon app initialization

When the app begins, the Screen1.Initialize event is triggered. The app calls the TinyDB1.GetValue with a tag of "responseMessage," the same tag you used when you stored the user's custom response entry earlier. The retrieved value is placed in the variable response so that it can be checked before we place it as the ResponseLabel. Can you think of why you'd want to check what you get back from the database before displaying it as the custom message to the user? TinyDB returns empty text if there is no data for a particular tag in the database. There won't be data the first time the app is launched; this will be the case until the user enters a custom response. Because the variable response now holds the returned value, we can use the if block to check if the length of what was returned by the database is greater than 0. If the length of the value contained in response is greater than 0, the app knows that TinyDB did return something, and the retrieved value can be placed into the ResponseLabel. If the length isn’t greater than 0, the app knows there is no previously stored response, so it doesn’t modify the ResponseLabel (leaving the default response in it).

Test your app.You cannot test this behavior through live testing, as the database gets emptied each time you "Connect AI Companion" to restart the app.

Instead, select "Build" then "App (provide QR code)" scan the barcode, and then download the app to your phone. Once the app is installed, enter a new response message in the NewResponseTextbox and click the SubmitResponseButton. Then close the app and restart it. Does your custom message appear?

Speaking the Incoming Texts Aloud

In this section, you'll modify the app so that when you receive a text, the sender's phone number, along with the message, is spoken aloud. The idea here is that when you’re driving and hear a text come in, you might be tempted to check the text even if you know the app is sending an autoresponse. With text-to-speech, you can hear the incoming texts and keep your hands on the wheel.

Android devices provide text-to-speech capabilities and App Inventor provides a component, TextToSpeech, that will speak any text you give it. (Note that here "text" is meant in the general sense of the word-a sequence of letters, digits, and punctuation-not an SMS text.)
In the "Getting Started" section of this app, we asked you to download a text-to-speech module from the Android Market. If you didn’t do so then, you’ll need to now. Once that module is installed and configured as desired, you can use the TextToSpeech component within App Inventor.

The TextToSpeech component is very simple to use-you just call its Speak function and plug in the text you want spoken into its message slot. For instance, the function shown in Figure 4-7 would say, "Hello World."
Figure 4-7. Blocks for speaking "Hello World" aloud

For the No Texting While Driving app, you'll need to provide a more complicated message to be spoken, one that includes both the text received and the phone number of the person who sent it. Instead of plugging in a static text object like the "Hello World" text block, you'll plug in a join block. An incredibly useful function, joinallows you to combine separate pieces of text (or numbers and other characters) into a single text object.

You'll need to make the call to TextToSpeech.Speak within the Texting.MessageReceived event handler you programmed earlier. The blocks you programmed previously handle this event by setting the PhoneNumber and Message properties of the Texting component appropriately and then sending the response text. You'll extend that event handler by adding the blocks listed in Table 4-6.

Table 4-6. Blocks for speaking the incoming text aloud

How the blocks work

After the response is sent, the TextToSpeech1.Speak function is called, as shown at the bottom of Figure 4-8. You can plug any text object into the message slot of the TextToSpeech1.Speak function. In this case,join is used to build the words to be spoken-it concatenates (or adds) together the text "SMS text received from" and the phone number from which the message was received (value number), plus the text ".The message is," and finally the message received (value messageText). So, if the text "hello" was sent from the number "111-2222," the phone would say, "SMS text received from 111-2222. The message is hello."

Figure 4-8. Speaking the incoming text aloud

Test your app.You’ll need a second phone to test your app. From the second phone, send a text to the phone running the app. Does the phone running the app speak the text aloud? Does it still send an automated response?

Adding Location Information to the Response

Apps like Facebook's Place and Google's Latitude use GPS information to help people track one another’s location. There are major privacy concerns with such apps, one reason being that location tracking kindles people's fear of a "Big Brother" apparatus that a totalitarian government might set up to track its citizens' whereabouts. But apps that use location information can be quite useful-think of a lost child, or hikers who've gotten off the trail in the woods.

In the No Texting While Driving app, location tracking can be used to convey a bit more information in response to incoming texts. Instead of just "I’m driving," the response message can be something like "I'm driving and I'm at 3413 Cherry Avenue." For someone awaiting the arrival of a friend or family member, this extra information can be helpful.

App Inventor provides the LocationSensor component for interfacing with the phone's GPS (or geographical positioning system). Besides latitude and longitude information, the LocationSensor can also tap into Google Maps to provide the driver's current street address.
It's important to note that LocationSensor doesn't always have a reading. For this reason, you need to take care to use the component properly. Specifically, your app should respond to the LocationSensor.LocationChanged event handler. 

LocationChanged event occurs when the phone’s location sensor first gets a reading, and when the phone is moved to generate a new reading. Using the blocks listed in Table 4-7, our scheme will respond to the LocationChanged event by placing the current address in a variable we'll name lastKnownLocation. Later, we'll change the response message to incorporate the address we get from this variable.

Table 4-7. Blocks to set up the location sensor

How the blocks work

The LocationSensor1.LocationChanged event is triggered the first time the sensor gets a location reading and when the device is moved so that a new reading is generated. Since you eventually want to send a street address as part of the response message, Figure 4-9 shows how the LocationSensor1.CurrentAddress function is called to get that information and store it in the lastKnownLocation variable. Behind the scenes, this function makes a call to Google Maps (via an API, something you’ll learn about in Chapter 24) to determine the closest street address for the latitude and longitude that the sensor reads.
Figure 4-9. Recording the phone's location in a variable each time the GPS location is sensed

Note that with these blocks, you've finished only half of the job. The app still needs to incorporate the location information into the autoresponse text that will be sent back to the sender. Let's do that next.

Sending the Location As Part of the Response

Using the variable lastKnownLocation, you can modify the Texting1.MessageReceived event handler to add location information to the response. Table 4-8 lists the blocks you'll need for this.

Table 4-8. Blocks to display location information in the autoresponse

How the blocks work


This behavior works in concert with the LocationSensor1.LocationChanged event and the variable lastKnownLocation. As you can see in Figure 4-10, instead of directly sending a message containing the text in ResponseLabel.Text, the app first builds a message using make text. It combines the response text in ResponseLabel.Text with the text “My last known location is:” followed by the variable lastKnownLocation.
Figure 4-10. Including location information in the response text

The default value of lastKnownLocation is "unknown," so if the location sensor hasn't yet generated a reading, the second part of the response message will contain the text "My last known location is: unknown." If there has been a reading, the second part of the response will be something like "My last known location is: 876 Willard Street, San Francisco, CA 95422."

Test your app.p. From the second phone, send a text to the phone running the app. Does the second phone receive the response text with the location information? If it doesn't, make sure you've turned GPS on in the first phone's Location settings.

The Complete App: No Texting While Driving


Figure 4-11 shows the final block configuration for No Texting While Driving.

Figure 4-11. The complete No Texting While Driving app (with all comments showing)

Variations

Once you get the app working, you might want to explore some variations. For example:

⦁ Write a version that lets the user define custom responses for particular incoming phone numbers. You'll need to add conditional (if) blocks that check for those numbers. For more information on conditional blocks, see Chapter 18.
⦁ Write a version that sends custom responses based on whether the user is within certain latitude/longitude boundaries. So, if the app determines that you're in room 222, it will send back "Bob is in room 222 and can’t text right now." For more information on the LocationSensor and determining boundaries, see Chapter 23.
⦁ Write a version that sounds an alarm when a text is received from a number in a "notify" list. For help working with lists, see Chapter 19.

Summary

Here are some of the concepts we've covered in this tutorial:

⦁ The Texting component can be used to both send text messages and process the ones that are received. Before calling Texting.SendMessage, you should set the PhoneNumber and Message properties of the Textingcomponent. To respond to an incoming text, program the Texting.MessageReceived handler.
⦁ The TinyDB component is used to store information persistently—in the phone's database-so that the data can be reloaded each time the app is opened. For more information on TinyDB, see Chapter 22.
⦁ The TextToSpeech component takes any text object and speaks it aloud.
⦁ make text is used to piece together (or concatenate) separate text items into a single text object.
⦁ The LocationSensor component can report the phone's latitude, longitude, and current street address. To ensure that it has a reading, you should access its data within the LocationSensor.LocationChanged event handler, which is triggered the first time a reading is made and upon every change thereafter. For more information on the LocationSensor, see Chapter 23.

If you're interested in exploring SMS-processing apps further, check out the Broadcast Hub app in Chapter 11.



댓글 없음:

댓글 쓰기